Passing User Parameter when invoking lambda in CodePipeline

I have a simple Lambda function that accepts bucket name as a parameter and displays the name of the bucket when the function is executed

import boto3

def lambda_handler(event, context):
    s3 = boto3.resource('s3')
    
    bucket_name = event['bucket_key']
    return bucket_name

If i pass this event JSON and execute it

{
  "bucket_key": "my_bucket"
}

I would get my_bucket as the response and it runs succesfully.

Now, what i want to do is to add this function in one of the stages as an action (Invoke Lambda) of my pipeline as shown below

However, when i run the pipeline, the function will go in a loop and if i look at cloudwatch, i can see this error message

[ERROR] KeyError: 'bucket_key'
Traceback (most recent call last):
  File "/var/task/lambda_function.py", line 6, in lambda_handler
    bucket_name = event['bucket_key']

Is there a specific format that i need to pass as the user parameter in codepipeline when invoking Lambda? It doesnt make sense to me that I'm passing exactly the same JSON event but it only works in lambda console

The code is missing a logic that will format the event received

import boto3
import json

pipeline = boto3.client('codepipeline')
s3 = boto3.resource('s3')

def lambda_handler(event, context):
    print(event) # to see what the event sent by the codepipeline
    user_parameters = event.get('CodePipeline.job', {}).get('data', {}).get('actionConfiguration', {}).get('configuration', {}).get('UserParameters', {})
    json_object = json.loads(user_parameters)
    print(json_object) # to see what response looks like after being formatted
    bucket_name = json_object.get('bucket_key',{}) # getting the bucket name
    index = 'index.html'

# Do some logic 
    try:
    # If bucket or file exist then we will execute the try block 
    s3.Object(bucket_name, index).load()
    print(f"{bucket_name} bucket and {index} exist! Taking down the frontend now ...")
    
    # Rename index to index.original
    s3.Object(bucket_name,'index-original.html').copy_from(CopySource=bucket_name+'/index.html')
    
    # Index-downtime overwrites index 
    s3.Object(bucket_name,'index.html').copy_from(CopySource=bucket_name+'/index-downtime.html')
    
    # Rename index original to index downtime
    s3.Object(bucket_name,'index-downtime.html').copy_from(CopySource=bucket_name+'/index-original.html')
    
    # Delete index-original.html 
    s3.Object(bucket_name,'index-original.html').delete()
    except:
        print(f"{bucket_name} bucket or {index} does not exist!")

    response = pipeline.put_job_success_result(
        jobId=event['CodePipeline.job']['id'] # needed for whenever codepipeline is invoking lambda
    ) 
    return response

Dont forget to add the put job success in lambda role